LLM Integration

Flint Pay publishes a machine-readable API reference at /SKILL.md, a single plain-text file designed to be pasted into an AI assistant's context window (CLAUDE.md, .cursorrules, system prompts, etc.).

If you need a structured schema instead of prose, Flint also publishes the public OpenAPI spec at /v1/openapi.json.

What's in SKILL.md#

The file is a self-contained markdown document covering:

  • Auth & base URL: the public API host plus test/live auth conventions
  • Public HTTP API surface: every documented /v1 endpoint, field, enum, and response shape
  • Conventions: money fields, ID prefixes, pagination, sparse updates and patch objects, metadata limits
  • State machines: order, payment, invoice, and subscription lifecycles
  • Error codes: structured error format with all domain-specific codes
  • Quick start examples: 5 curl commands covering the most common flows

The file stays under 2k lines, so an agent can read it in one pass instead of navigating rendered doc pages, and it gives the agent correct field names, valid state transitions, and working curl commands to adapt rather than hallucinated fields and guessed syntax.

The SKILL.md file tracks the same API surface as these docs. When new endpoints or fields ship, both are updated together.

How to Use It#

Add a WebFetch instruction to your project's CLAUDE.md:

markdown
# Flint Pay API

When working with the Flint Pay API, fetch the reference:
https://developers.withflintpay.com/SKILL.md

Claude Code will fetch and read the file when it needs API context.

https://developers.withflintpay.com/SKILL.md

Letting an Agent Call the API#

SKILL.md and the MCP server give an agent knowledge. To let it act, the Flint CLI is the shortest path: it exposes the whole command surface as machine-readable JSON Schema, so an agent can discover a command, build a valid request body, and validate it before spending a call.

Bash
flint schema commands --output json
flint schema input payment-intents.create --output json

--output json also forces non-interactive mode, so no command blocks on a prompt. Agents that speak MCP can skip the shell entirely with flint mcp serve, which serves the same commands as MCP tools over stdio.

Rate this doc